[1] https://www.sciencedaily.com/releases/2015/03/150317122458.h...
[2] https://pubmed.ncbi.nlm.nih.gov/28768407/ (two years later)
"... in a correct analysis of the data used by the Institute of Medicine, it was found that 8895 IU/d was needed for 97.5% of individuals to achieve values ≥50 nmol/L."
That just means there are some people who need a very high dose. For comparison, 4000IU/d gets me well over 50nmol/L, this is also the "generally considered safe" upper limit. There was a case reported on HN where 5000IU/daily caused issues with calcium levels. You gotta get your bloodwork done.
In my personal example, taking about 6000 UI (plus 300 mcg K2) for a year put me just under 200 nmol/l. I did an experiment and stopped taking it for 2 months (but kept taking K2). After testing again I had dropped to 70 nmol/l (top of the insufficient range) and had mild symptoms of muscle ache and fatigue. Needless to say, I've started taking it again (and will aim for 175 nmol/l).